/* * netsniff-ng - the packet sniffing beast * Copyright 2012 Daniel Borkmann. * Subject to the GPL, version 2. * Parts derived from iw, subject to ISC license. * Copyright 2007, 2008 Johannes Berg * Copyright 2007 Andy Lutomirski * Copyright 2007 Mike Kershaw * Copyright 2008-2009 Luis R. Rodriguez */ #define _GNU_SOURCE #include #include #include #include #include #include #include #include #include #include #include #include #include #include "die.h" #include "str.h" #include "dev.h" #include "mac80211.h" #include "xmalloc.h" #include "built_in.h" struct nl80211_state { struct nl_sock *nl_sock; struct nl_cache *nl_cache; struct genl_family *nl80211; }; static void get_mac80211_phydev(const char *device, char *phydev_path, size_t phydev_len) { int ret; char *pathstr; ssize_t num; ret = asprintf(&pathstr, "/sys/class/net/%s/phy80211", device); if (ret < 0) panic("Can't generate path name string for /sys/class/net device"); num = readlink(pathstr, phydev_path, phydev_len); if (num < 0) { if (errno == ENOENT || errno == EINVAL) panic("It's probably not a mac80211 device!\n"); panic("Can't readlink %s: %s!\n", pathstr, strerror(errno)); } xfree(pathstr); phydev_path[min_t(size_t, num, phydev_len - 1)] = 0; } static inline struct nl_msg *nl80211_nlmsg_xalloc(void) { struct nl_msg *ret = nlmsg_alloc(); if (!ret) panic("Cannot allocate nlmsg memory!\n"); return ret; } static inline struct nl_sock *nl80211_nl_socket_xalloc(void) { struct nl_sock *ret = nl_socket_alloc(); if (!ret) panic("Cannot allocate nl socket memory!\n"); return ret; } static void nl80211_init(struct nl80211_state *state) { int ret; state->nl_sock = nl80211_nl_socket_xalloc(); ret = genl_connect(state->nl_sock); if (ret) panic("Cannot connect generic netlink!\n"); ret = genl_ctrl_alloc_cache(state->nl_sock, &state->nl_cache); if (ret < 0) panic("Failed to allocate generic netlink cache: %s!", nl_geterror(-ret)); state->nl80211 = genl_ctrl_search_by_name(state->nl_cache, "nl80211"); if (!state->nl80211) panic("nl80211 not found in netlink cache!\n"); } static void nl80211_cleanup(struct nl80211_state *state) { genl_family_put(state->nl80211); nl_cache_free(state->nl_cache); nl_socket_free(state->nl_sock); } static int nl80211_wait_handler(struct nl_msg *msg __maybe_unused, void *arg) { int *finished = arg; *finished = 1; return NL_STOP; } static int nl80211_error_handler(struct sockaddr_nl *nla __maybe_unused, struct nlmsgerr *err, void *arg __maybe_unused) { panic("nl80211 returned with error (%d): %s\n", err->error, nl_geterror(err->error)); } static int nl80211_add_mon_if(struct nl80211_state *state, const char *device, const char *mondevice) { int ifindex, ret; struct nl_msg *msg; struct nl_cb *cb = NULL; int finished = 0; ifindex = device_ifindex(device); msg = nl80211_nlmsg_xalloc(); genlmsg_put(msg, 0, 0, genl_family_get_id(state->nl80211), 0, 0, NL80211_CMD_NEW_INTERFACE, 0); NLA_PUT_U32(msg, NL80211_ATTR_IFINDEX, ifindex); NLA_PUT_STRING(msg, NL80211_ATTR_IFNAME, mondevice); NLA_PUT_U32(msg, NL80211_ATTR_IFTYPE, NL80211_IFTYPE_MONITOR); ret = nl_send_auto_complete(state->nl_sock, msg); if (ret < 0) { if (ret == -ENFILE) { nlmsg_free(msg); return -EBUSY; } panic("Cannot send_auto_complete!\n"); } cb = nl_cb_alloc(NL_CB_CUSTOM); if (!cb) panic("Cannot alloc nl_cb!\n"); nl_cb_set(cb, NL_CB_ACK, NL_CB_CUSTOM, nl80211_wait_handler, &finished); nl_cb_err(cb, NL_CB_CUSTOM, nl80211_error_handler, NULL); nl_recvmsgs(state->nl_sock, cb); if (!finished) { ret = nl_wait_for_ack(state->nl_sock); if (ret < 0) { if (ret == -ENFILE) { nlmsg_free(msg); return -EBUSY; } panic("Waiting for netlink ack failed!\n"); } } nl_cb_put(cb); nlmsg_free(msg); return 0; nla_put_failure: panic("nla put failure!\n"); return -EIO; /* dummy */ } static int nl80211_del_mon_if(struct nl80211_state *state, const char *mondevice) { int ifindex, ret; struct nl_msg *msg; ifindex = device_ifindex(mondevice); msg = nl80211_nlmsg_xalloc(); genlmsg_put(msg, 0, 0, genl_family_get_id(state->nl80211), 0, 0, NL80211_CMD_DEL_INTERFACE, 0); NLA_PUT_U32(msg, NL80211_ATTR_IFINDEX, ifindex); ret = nl_send_auto_complete(state->nl_sock, msg); if (ret < 0) panic("Cannot send_auto_complete!\n"); ret = nl_wait_for_ack(state->nl_sock); if (ret < 0) panic("Waiting for netlink ack failed!\n"); nlmsg_free(msg); return 0; nla_put_failure: panic("nla put failure!\n"); return -EIO; /* dummy */ } void enter_rfmon_mac80211(const char *device, char **mondev) { int ret; short flags; uint32_t n; char phydev_path[256]; struct nl80211_state nlstate; /* XXX: is this already a monN device? */ get_mac80211_phydev(device, phydev_path, sizeof(phydev_path)); nl80211_init(&nlstate); for (n = 0; n < UINT_MAX; n++) { char mondevice[32]; slprintf(mondevice, sizeof(mondevice), "mon%u", n); if (device_ifindex_get(mondevice) > 0) continue; ret = nl80211_add_mon_if(&nlstate, device, mondevice); if (ret == 0) { *mondev = xstrdup(mondevice); flags = device_get_flags(*mondev); flags |= IFF_UP | IFF_RUNNING; device_set_flags(*mondev, flags); nl80211_cleanup(&nlstate); return; } } panic("No free monN interfaces!\n"); } void leave_rfmon_mac80211(const char *mondev) { short flags; struct nl80211_state nlstate; flags = device_get_flags(mondev); flags &= ~(IFF_UP | IFF_RUNNING); device_set_flags(mondev, flags); nl80211_init(&nlstate); nl80211_del_mon_if(&nlstate, mondev); nl80211_cleanup(&nlstate); } 421a4ca76468'>Documentation/vm/numa parent0b3589be9b98994ce3d5aeca52445d1f5627c4ba (diff)
perf/x86/intel/rapl: Make package handling more robust
The package management code in RAPL relies on package mapping being available before a CPU is started. This changed with: 9d85eb9119f4 ("x86/smpboot: Make logical package management more robust") because the ACPI/BIOS information turned out to be unreliable, but that left RAPL in broken state. This was not noticed because on a regular boot all CPUs are online before RAPL is initialized. A possible fix would be to reintroduce the mess which allocates a package data structure in CPU prepare and when it turns out to already exist in starting throw it away later in the CPU online callback. But that's a horrible hack and not required at all because RAPL becomes functional for perf only in the CPU online callback. That's correct because user space is not yet informed about the CPU being onlined, so nothing caan rely on RAPL being available on that particular CPU. Move the allocation to the CPU online callback and simplify the hotplug handling. At this point the package mapping is established and correct. This also adds a missing check for available package data in the event_init() function. Reported-by: Yasuaki Ishimatsu <yasu.isimatu@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Sebastian Siewior <bigeasy@linutronix.de> Cc: Stephane Eranian <eranian@google.com> Cc: Vince Weaver <vincent.weaver@maine.edu> Fixes: 9d85eb9119f4 ("x86/smpboot: Make logical package management more robust") Link: http://lkml.kernel.org/r/20170131230141.212593966@linutronix.de Signed-off-by: Ingo Molnar <mingo@kernel.org>
Diffstat (limited to 'Documentation/vm/numa')